Ruicheng Wang

> I think an async version would be great. Unfortunately I do not have time at the moment to dedicate to converting it to async, but pull requests or any...

> Do you mean the line below? It seems we have already set it as `True`. > > https://github.com/LiheYoung/Depth-Anything/blob/f419b7db90b26b2855280c4da484778c4fac759f/metric_depth/zoedepth/models/zoedepth/config_zoedepth.json#L50 Oh, I see. It seems that `infer` mode has `force_keep_ar` set...

Sorry that I didn't make it clear. I just checked the original repo. The evaluation configuration of ZoeDepth in the *original repo* is the same with depth-anything (with `force_keep_ar=false` in...

每次要跑一个依赖 mmcv 的项目我就会来巡回一下

Hi! I'm glad to hear it was helpful. Regarding the scale consistency across multiple inferences, a simple solution is to use **MoGe-2**, which supports **metric-scale prediction**. It should provide consistent...

Ah, I just noticed that the current script for merging panorama depth maps doesn't account for metric scale. It ignores input metric scales. The merged panorama depth needs to be...

Thanks for bringing up this idea. it definitely makes sense. In our current pipeline, we only correct radar–camera synchronization errors during dataset processing by using the ego-vehicle speed. This is...

Hi, thank you for your interest! Your idea sounds solid and straightforward! We could consider extending the self-attention mechanism or adding cross-attention layers in the ViT to enable multi-image inputs...

Hi! The `model.infer()` function supports both single-image inputs of shape (3, H, W) and batches of images (B, 3, H, W). The optional `fov_x` argument also supports both scalar and...

Hi! Sorry for the late response. We've been focusing on cleaning up the training code and improving the model's performance over the past few months. The training code, along with...